Maria Cantwell is the Democratic senior senator from Washington and the ranking member of the Senate Committee on Commerce, Science, and Transportation, the committee with jurisdiction over the frontier AI legislation under negotiation in 2026.
Position on frontier-model review
Cantwell has been contemplating a pre-deployment regime under which the government would vet frontier models before release, out of a stated concern that developers are not best positioned to test their own systems. That position is the counterweight to the duty-of-care approach favored by Majority Leader John Thune, Sen. Amy Klobuchar and Commerce chair Ted Cruz, a split that became public on August 3, 2026; Cantwell's spokesperson declined to comment on the account (Source: washingtonpost.com).
At a July 30, 2026 Senate Commerce telecommunications subcommittee hearing she pressed for mandatory guardrails and rejected industry self-certification, citing cyber and biosecurity risks and pointing to the disclosure of the prior week: "We saw last week with OpenAI's disclosure of an unprecedented security incident — one of its advanced models broke out of the sandbox and created a complex cyberattack against Hugging Face" (Source: insideaipolicy.com). She restated the position in remarks that became public on August 4, 2026, again emphasizing cyber and biosecurity imperatives for mandatory guardrails and dismissing self-certification, at a subcommittee hearing on AI and telecom systems; the article body is paywalled and the hearing date was not retrievable from the lede (Source: insideaipolicy.com).
By September 2026 she had specified where she would place the testing function. Writing on X as the Thune–Klobuchar–Cruz negotiation came to a head, she said that "meaningful legislation would require the most powerful AI models undergo testing by scientists and experts at our national laboratories to assess whether they could enable sophisticated cyberattacks or aid the development of biological or nuclear weapons," and separately posted "I welcome the urgency" (Source: reuters.com; nextgov.com).
Role in the 2026 Senate negotiation
Cantwell is not a principal in the Thune–Klobuchar–Cruz talks but, as ranking Democrat on the committee of jurisdiction, has weighed in on the draft (Source: reuters.com). The element she objects to, according to a Democratic committee aide, is a regime in which safety tests are conducted by the companies themselves and the results presented to the Commerce secretary for deployment approval; the same aide described the Cruz–Klobuchar–Thune version she opposes as "primarily a voluntary standard type situation," and said Cantwell's office had offered alternative language to Klobuchar's office, which was "very receptive," with Cruz's approval the deciding factor (Source: nextgov.com). Klobuchar was reported to be wary of introducing the draft over Cantwell's objection, which is one reason the September 23, 2026 markup Cruz sought remained unsettled; two people familiar with the plans said Cantwell "has major concerns with the proposed framework" (Source: washingtonpost.com; washingtonpost.com).
On September 15, 2026 an account described Cantwell as not then part of the negotiations, having given feedback in the past; she said on the Senate floor, "We need our colleagues to say, 'Stop with saying the industry can do what it wants' and let's put together the infrastructure at the federal level that not only has strong federal standards, but also has independent testing and real safeguards for the American people" (Source: rollcall.com). The next day she said she was working with Klobuchar on the bill: "We think it's part of the equation, we don't think it's the whole story" (Source: semafor.com).
Thune said on September 23, 2026 that Cantwell was asking for AI rulemaking authority at the Department of Commerce, one of Democrats' two main sticking points on the Thune–Klobuchar bill, and that he worries such authority could give the executive too much power; Klobuchar was seeking Cantwell as a co-sponsor before reintroducing the bill (Source: axios.com).
JAWBONE Act
At the Commerce Committee markup of September 17, 2026, Cantwell said she supports the goals of Cruz's JAWBONE Act but argued that, with Congress having enacted no AI regulation, the text would let "almost any government conversation with an AI company [about] safety" be called coercion. Her example was a Department of Energy employee asking a company to stop its model from explaining how to enrich uranium, which she said "could be illegal under this bill, and the employee could be sued." She said she wanted national-laboratory scientists to work with AI companies on risks such as step-by-step instructions for making pathogens like Ebola, and asked the committee "not to have a chilling effect on the communication at an urgent time." Her substitute amendment would have removed AI models from the bill, struck the public portal of government communications, limited liability to decision-making officials, and required a judge to find any consent judgment in the public interest. It was not agreed to; Cruz committed to work with her to address the concerns (Source: commerce.senate.gov).
Future of AI Innovation Act
Cantwell is an original cosponsor of S. 3952, the Future of Artificial Intelligence Innovation Act of 2026, introduced on February 26, 2026 by Sen. Todd Young (R-Indiana) with Cantwell, Sen. Marsha Blackburn and Sen. John Hickenlooper (Future of Artificial Intelligence Innovation Act of 2026 (S. 3952)). It is a reintroduction of S. 4178 of the 118th Congress, which Cantwell sponsored in April 2024 with Young, Hickenlooper and Blackburn as original cosponsors, and which the Commerce Committee ordered reported on July 31, 2024 and Cantwell reported to the Senate on December 18, 2024 (Source: congress.gov). The Commerce Committee minority's release on the 2026 reintroduction quotes her: "This legislation brings together private sector and government experts to develop voluntary standards for AI, create new assessment tools, and conduct testing that will ensure the United States leads in AI-driven innovation and competitiveness for decades to come," and "The NIST Center for AI Standards and Innovation, testbeds at our national labs, and grand challenge prizes will accelerate new advancements while helping companies and consumers better use AI in a safe and secure manner" (Source: commerce.senate.gov).
The bill would formally establish the Center for AI Standards and Innovation within NIST as a new section 22B of the NIST Act, with a rule of construction that "Nothing in this section shall be construed to provide the Director any enforcement authority that was not in effect on the day before the date of the enactment of the Future of Artificial Intelligence Innovation Act of 2026." It would also create a testbed program led jointly by NIST and the Department of Energy, in coordination with the National Science Foundation, that draws on the National Laboratories and includes a program giving foundation-model vendors "the opportunity to voluntarily test foundation models" (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)). The Washington Post described the bill as one that would formally establish the Center and create public-private testbeds at the Department of Energy, the National Science Foundation and NIST, and reported that the roster of the September 23, 2026 markup Cruz sought would likely include it, according to one person familiar with the committee's plans (Source: washingtonpost.com; washingtonpost.com).
The bill relates to her testing position in two ways. It would give a statutory basis to the Center and place an AI testing program at the national laboratories, where she has said the most powerful models should be tested. Its testing provisions are voluntary (Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026)), whereas her September 2026 statement called for legislation that "would require the most powerful AI models undergo testing" there (Source: reuters.com).
Open questions
- Whether the alternative testing language Cantwell's office offered was incorporated into the Klobuchar–Thune draft is not recorded.
Relationships
- related: John Thune, Amy Klobuchar, Ted Cruz (the competing duty-of-care position); AI Pre-Release Vetting (the instrument she favors); AI and Tort Liability; OpenAI and Hugging Face Partner to Address Security Incident During Model Evaluation (OpenAI, July 2026) (the incident she cites)
- contradicts: John Thune — the disagreement is over who runs the tests, not over whether the most capable models should be tested; both positions are inside one negotiation and neither has prevailed
- supports: NIST CAISI (Center for AI Standards and Innovation) — the Future of AI Innovation Act she cosponsors would give the Center a statutory basis, alongside a joint NIST–DOE testbed program coordinated with NSF
- related: Future of Artificial Intelligence Innovation Act of 2026 (S. 3952), Future of Artificial Intelligence Innovation Act of 2026 — S. 3952 bill text (introduced February 26, 2026) — original cosponsor of S. 3952 (2026); sponsor of the predecessor S. 4178 (2024)
- related: Todd Young, Marsha Blackburn, John Hickenlooper — co-sponsors of S. 3952
- related: JAWBONE Act — her September 17, 2026 substitute amendment would have removed AI models from the bill